Josh Triplett

This looks correct to me.

I do think we're likely to want deregistration as well, but I think we'd want to do that via a scoped function rather than naming. I don't think it *has*...

I think hooks are going to be quite prone to potential deadlocks. That's not a blocker for doing this, just something that'll need to be documented.

Rather than treating this purely as diagnostic, what if we treated it as directly actionable metadata? If someone tries to install a crate, and it isn't installable because it's a...

@epage Not following how this would go circular. It would only lead cargo from attempting to install the lib to actually installing the bin, and installing the bin would likely...

@epage I was thinking of it as a hint, which doesn't need validating. And I don't think it needs versioning.

@epage Installing the latest version doesn't seem *worse* than completely failing. Or, alternatively, we could skip automatically using `recommended-bin-packages` if you specify an `@version`, and instead fail with an *error*...

@lopopolo I think that would be a separate larger change, since bstr is built around `u8`. It'd be doable but would involve more code.

The CI failure seems unrelated. https://github.com/BurntSushi/bstr/pull/192 fixes it.

In this context, though, I think we don't need the structure to have a size.